critical obstacle to molecular and genomic classification of primary melanoma has been the requirement to provide prognostic measures (i.e. Breslow thickness) and related histologic parameters in formalin-fixed, paraffin-embedded (FFPE) biospecimens, where unlike frozen material, unadulterated mRNA and DNA have been difficult or impossible to extract and study. With our collaborators, we have successfully applied a novel technology first developed by Fan et al in 2004 for global gene expression profiling of melanoma FFPE biospecimens. This platform, designated DASL for DNA-mediated Annealing, Selection, extension, and Ligation) renders formalin-fixed biospecimens embedded and archived in the form of paraffin blocks usable for gene expression analysis. We found two distinct subclasses characterized by two robust signatures: one broadly defined by a cluster of microphthalmia and melanocyte differentiation genes ("MITF" class) and a more heterogeneous group of inflammation, innate immune, and growth-related genes ("IIG" class). Our preliminary data indicate that DASL can successfully profile and recover the 2-class structure from FFPE primary melanomas in the Nurses Health Study 1. The Channing Laboratory at Brigham and Women's Hospital is home to large cohort studies, i.e. the Nurses'Health Study 1 (NHS1), Nurses'Health Study-2 (NHS2) and the Health Professionals Follow-up Study (HPFS). The further strength of these cohorts lies in disease follow-up with bi-annual updates on new cases and mortality and availability of tissue blocks from the primary tumor site for melanoma cases in the cohorts. We hypothesize that melanoma aggressiveness is determined by MITF/IIG classes and we propose to evaluate whether this 2-class structure is associated with known histological parameters (e.g. Breslow thickness), and melanoma recurrence/mortality. With access to unique population-based national cohorts and development of new technologies with our collaborators, the specific aims of this proposal are as follows: Specific Aim 1: Develop molecular signatures of NHS, NHS2 and HPFS melanoma FFPE samples using an Illumina expression array platform with over 6000 genes and a Sequonome platform to sequence MC1R and genotype specific loci for BRAF and NRAS to: a. Assign MITF or IIG class for each melanoma sample and evaluate their association with BRAF/NRAS/MC1R status and determinants of melanoma histology (e.g. Breslow thickness, Clark's level, vertical growth phase) b. Evaluate genome-wide (6,000 genes) class discovery for each melanoma sample and evaluate their association with BRAF/ NRAS/MC1R status and determinants of melanoma histology (e.g. Breslow, Clark's level, vertical growth phase) Specific Aim 2: Using the transcriptional and genomic signatures from Aim 1 to: a. Assign MITF or IIG class association with BRAF/NRAS/MC1R status and melanoma recurrence and mortality due to melanoma b. Evaluate genome-wide (6,000 genes) association with BRAF/NRAS/MC1R status and melanoma recurrence and mortality due to melanoma PUBLIC HEALTH RELEVANCE: Malignant melanoma of the skin is a lethal cancer with significant impact on morbidity and mortality. The clinical paradox is that melanoma is curable when diagnosed early and quite lethal with advanced disease. Since validation of the Clark's level for melanoma invasiveness in the 1960's, there has been little change in our understanding of the significant association between histologic grading of melanoma and melanoma recurrence and mortality. We hope to be able to identify biomarkers that point to pathways of metastatic potential and assist in earlier prognostication. Thus we aim to significantly improve the current paradigm for diagnosis and therapy for melanoma by eventually reducing mortality from this deadly skin cancer.